John Smythies (1999) The Biochemical Basis of Coma. Psycoloquy: 10(026) Coma Biochemistry (1)

Volume: 10 (next, prev) Issue: 026 (next, prev) Article: 1 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 10(026): The Biochemical Basis of Coma

Target Article on Coma-Biochemistry

John Smythies
Division of Neurochemistry
Brain and Perception Laboratory
Center for Brain and Cognition
University of California, San Diego
La Jolla CA 92093-0109
Department of Neuropsychiatry
Institute of Neurology
Queen Square, London


Current research on the neural basis of consciousness is based mainly on neuroimaging, physiology and psychophysics. This target article reviews what is known about biochemical factors that may contribute to the development of consciousness, based on loss of consciousness (i.e., coma). There are two theories of the biochemical mode of action of general anaesthetics. One is that anaesthesia is a direct (i.e., not receptor-mediated) effect of the anaesthetic on cellular neurophysiological function; the other is that some alteration of receptor function occurs. General anaesthetics are mainly GABA agonists but some (such as ketamine) are glutamate antagonists. They also affect other systems, particularly cholinergic ones. There are various comas of metabolic origin. For example, a combination of small doses of the iron chelators desferrioxamine and prochlorperazine induce a profound and long lasting coma in humans. The mechanisms that might mediate this include redox mechanisms at the glutamate synapse, post-synaptic endocytosis of dopamine and iron, and intracellular iron-dopamine complexes, which are powerful dismuters of the superoxide anion. New findings in cell biology relating to endocytosis and recycling of receptors are discussed in a wider context. These biochemical events may induce coma by two mechanisms: (i) Consciousness may depend on widespread cortical (or cortico-thalamic) activation. (ii) Whereas these biochemical changes are widespread, only the changes in a subset of 'consciousness' neurons may count. An experimental program to distinguish between these two alternatives is proposed.


anaesthetics, coma, consciousness, desferrioxamine dopamine, GABA, general glutamate iron, redox mechanisms, synapses
    The target article below has just appeared in PSYCOLOQUY, a
    refereed journal of Open Peer Commentary sponsored by the American
    Psychological Association. Qualified professional biobehavioral,
    neural or cognitive scientists are hereby invited to submit Open
    Peer Commentary on it. Please email or see websites for
    Instructions if you are not familiar with format or acceptance
    criteria for PSYCOLOQUY commentaries (all submissions are

    To submit articles and commentaries or to seek information:


    reviews the biochemical basis of various types of coma, including
    general anesthesia. This is a subject of considerable current interest
    in its own right in neuroscience and anesthesiology and may throw some
    light on the biochemical mechanisms of consciousness. It also concerns
    the interrelations among several biochemical and microanatomical
    systems whose investigators normally do not have much contact. This
    interdisciplinary communication could lead to interesting new
    developments in these diverse fields (e.g., concerning the role of
    redox reactions at the glutamate synapse, the role of endocytosis in
    postsynaptic functions, the role of dismuting catecholamine-iron
    complexes in the brain, etc.). Commentary would accordingly be welcome
    from specialists in the following fields: the neurochemistry of the
    glutamate receptor and of synaptic plasticity, reactive oxygen species
    and antioxidants, trafficking mechanisms for iron and possibly
    catecholamines at postsynaptic sites in the brain, the role of
    catecholamines as antioxidants (in particular in the case of their
    complexes with iron), oxidative pathways of catecholamine metabolism in
    the brain, the function of endocytosis in neurons, and the
    neuropharmacology of general anesthesia.


1. Francis Crick (1994) has suggested that one way of discovering the neural basis of consciousness is to identify (i) those neurons that are active during conscious states and inactive during states of loss of consciousness, and (ii) those neurons whose activity is unaffected by the state of consciousness, and then see if we can determine differences between the two populations. Given our increasing knowledge of the plasticity of neurons, it may also be possible that within a single population of neurons, one type of activity may be associated with conscious states and another with unconscious ones. Most current work stimulated by this strategy concentrates upon brain imaging that identifies the macroanatomical areas involved, on physiological studies that look at such factors as synchronization of 40 Hz rhythms, or on psychophysical investigations using such phenomena as retinal rivalry.

2. It cannot merely be assumed, however, that only the electrical activity of neurons is necessary and sufficient for the generation of the conscious state. It is theoretically possible that some of the relevant differences that Crick is looking for between 'consciousness' and 'non-consciousness' neurons may be biochemical. If so, it might be useful to consider what these biochemical factors might possibly be. The purpose of this review is to see whether any useful leads can be obtained by examining the biochemical changes which produce coma. This in turn may suggest mechanisms relevant to the conscious state, always making the Jacksonian caveat that, if a certain biochemical change A leads to loss of function X, one cannot assume that the normal function of A is to produce X. Nevertheless, a study of the biochemical mechanisms underlying coma, if cautiously interpreted, may throw some light on the sort of mechanisms to investigate during the conscious state.


3. There are two schools of thought with regard to the way anaesthetics work. The earlier theory is that these agents, which are lipophilic molecules, act by infiltrating the lipid plasma membrane and disrupting its function mechanically (Janoff et al, 1981). A more modern version of this classical 'general' theory (i.e., non-receptor mediated) has been presented by Reis and Puil (1999a, b). Using thalamic brain slices from rats, they showed that isoflurane hyperpolarizes thalamic neurons by increasing K+ conductance (leak). This inhibits neuronal firing by short circuiting the effectiveness of depolarizing pulses and shunting voltage dependent Na+ and Ca2+ channels. This effect still occurs in the presence of GABA receptor antagonists. The authors note that the K+ mechanism is ubiquitous in cells and that anaesthetics affect a number of other systems, but they also claim that the effect they report on the cortico-thalamic-cortical system could by itself account for the general anaesthesia. Anoxia (which also induces coma) hyperpolarizes CNS neurons by increasing K+ conductance, and in deep slow-wave sleep thalamo-cortical neurons hyperpolarize and show spike-bust rhythms, but the basic mechanism is different.

4. In anaesthesia there is an increase in Na+ -spike thresholds and in decrease in low threshold calcium spikes not seen in deep sleep. Furthermore, in deep sleep thalamo-cortical neurons hyperpolarize because of cholinergic disfacilitation leading to oscillations. In anaesthesia the hyperpolarization is due to the increase in K+ conductance and there is no concomitant shift in either tonic or burst firing. Reis and Puil (1999) conclude that their results fit in with the 'classical' (non-receptor mediated) theory of the action of anaesthetics. Further support for the 'specific neural network' (i.e., local) theory has recently been supplied by Fiset et al (1999). They carried out PET studies on normal volunteers given propofol at three doses. They report that anaesthesia is associated with a generalized decrease in cortical blood flow, but there is a preferential decrease in blood flow (bilateral) in the medial thalamus, cuneus and precuneus, and in the posterior cingulate and orbitofrontal cortices as well as in the right angular gyrus. These are all brain regions previously associated with arousal, associative functions and autonomic control. These authors conclude that their data supports the specific network theory of the action of general anaesthetics rather than the non-specific generalized theory.

5. The alternative theory is that general anaesthesia is mediated by local receptors. Most general anaesthetics (such as barbiturates and halothane) inhibit GABAergic transmission by enhancing the receptor binding of GABA and increasing the receptivity of the GABA receptor channel (Franks and Lieb, 1994). They probably bind directly to lipophilic pockets in the receptor protein (Eckenhoff and Johansson, 1997). There is evidence to suggest that they do not affect presynaptic transmitter release (Liachenko et al, 1998). However, the general anaesthetic ketamine, like nitrous oxide ('laughing gas'), is a glutamate NMDA receptor blocker, and acts differently (Franks and Lieb,1998). It modulates the action of GABA on the GABA receptor, but has no action by itself on this receptor (Dzoljic & van Duijn, 1998).

6. It would be naive to suggest that anaesthetics act by only one mechanism. The biochemistry of synaptic systems is enormously complicated and anaesthesia could affect a number of these.

    (i) Numerous studies have been made of the effect of general
    anaesthetics on the cholinergic systems. General anaesthetics have
    been shown to increase rates of desensitization at the nicotinic
    ACh receptor (Raines et al, 1995; Scheller et al, 1997; Liu et al,
    1995), and to act as antagonists (Andoh et al, 1997; Franks and
    Lieb, 1997). Ether increases the burst frequency at this receptor
    by promoting agonist binding to the receptor protein (Liu et al,
    1994).  These anaesthetics also inhibit muscarinic receptors
    (Minami et al, 1997). Keifer et al (1996) have suggested that a
    cholinergic mechanism generates the key midbrain pedunculopontine
    and lateral dorsal tegmental nuclei in EEG sleep spindles.
    Anaesthetic agents also block choline uptake into synaptosomes,
    which would inhibit cholinergic functions (Griffiths et al, 1994).
    Experiments using brain microdialysis volatile anaesthetics have
    shown that these anaesthetics suppress ACh release, while nitrous
    oxide has the opposite effect (Shichino et al, 1998). Some cases of
    hepatic coma have a cholinergic element, which, as with the coma
    produced by atropine, can be relieved by intravenous physostigmine
    (Kabatnik et al, 1999).

    (ii) Both halothane and ketamine inhibit Ca2+-activated K channels
    and their phospholipase A2-arachidonic acid signal transduction
    pathways (Denson et al, 1996). Both ketamine and halothane also
    inhibit K currents through Kv2-1 channels, but do so by different
    mechanisms (Kulkarni et al, 1996).

    (iii) Volatile anaesthetics selectively inhibit plasma membrane
    Ca2+-transport APTase by binding to lipophilic sites and changing
    conformation (Lopez & Kosk-Kosicka 1995).

    (iv) Halothane inhibits Na+-mediated glutamate release (Ratnakumari
    & Hemmings, 1998). Liachenko et al (1998) report that the
    anaesthetic cyclobutane derivative F3 inhibits K+-evoked glutamate
    and GABA release, whereas the non-anaesthetic cyclobutane
    derivative F6 suppresses evoked glutamate release but has no effect
    on evoked GABA release. They conclude that the suppression of
    excitatory neurotransmitter release might not be the mode of action
    of general anaesthetics. The suppression of GABA release is
    somewhat paradoxical.

    (v) In the molecular layer of the cerebellum NMDAr stimulation
    activates the nitric oxide-guanyl cyclase signaling pathway. This
    effect is inhibited by halothane and isoflurane (Zou et al, 1996).

    (vi) Halothane increases the probability of opening of the
    glycine-activated channel in rat central neurons (Wakamori et al,

    (vii) Isofluorane inhibits histamine metabolism in the hypothalamus
    (Hashimoto et al, 1998), and affects polyamine amine metabolism in
    the brain (Mills et al, 1997).

    (viii) Isoflurane disrupts synchronized neural oscillations at a
    frequency of <10 Hz (rather than ~40 Hz) (Tennigkeit et al, 1997).

    (ix) Low doses of anaesthetics induce hypethesis, i.e., disruption
    of memory formation without loss of awareness. On this basis,
    Andrade (1996) suggests that frontal lobe function is particularly
    sensitive to anaesthetics.

    (x) With regard to systems, Flohr (1995) suggests that
    consciousness involves the NMDAr mediated activation of large
    neuronal assemblies, and that anaesthetics and brain stem lesions
    have a common denominator for inducing coma, namely inhibition of
    the formation of these assemblies.

7. In conclusion, it appears that one important role for volatile anaesthetics is to potentiate GABA-mediated inhibition at the receptor level and to inhibit NMDAr-mediated conduction at the level of the post-synaptic signaling cascade (Zou et al, 1996). Ketamine-like agents inhibit glutamate-mediated conduction at the receptor level, while nitrous oxide affects both systems. There is also evidence that other systems, particularly the cholinergic, may be important.


8. Some comas of metabolic origin also involve the glutamate and/or GABA systems. For example, hepatic coma is associated with raised blood levels of GABA and ammonia, which inhibits metabotropic glutamate receptors (Albrecht, 1998). Abnormal sugar metabolism in comas suggests that other systems are involved. Diabetic coma has been linked to a reduction in brain cell volume caused by changes in brain osmolarity (Fink et al, 1994). However, in this instance there was also an increase of GABA release in the cortex. Hypoglycaemic coma may be related to increased tyrosine phosphorylation of mitogen-activated protein kinase (Kurihara and Wielock, 1994). Concussion has been linked to the excessive release of glutamate (Bullock et al, 1998), of oxygen free radicals, in particular superoxide anions (Mori et al, 1998; Yunoki et al, 1998), and to changes in brain osmolarity (Katayama et al, 1998).


9. In 1985, Blake et al reported that a combination of the standard iron-chelator desferrioxamine (100 mg) plus prochlorperazine (25 mg) produces a profound and prolonged coma lasting 48-72 hours in humans, whereas given singly, these drugs had no such effect. A similar result was obtained with rats. The effect was explained as follows. Desferrioxamine is a hydrophilic iron chelator and prochlorperazine is a lipophilic iron chelator. This drug combination was shown to act synergistically in transferring iron across a layer of chloroform between two water compartments. The researchers suggested that this combination produces a rapid flux of iron (and copper) out of neurons, which in turn produces a disturbance of the plasma membrane, leading to interference with serotoninergic and noradrenergic function, and coma.

10. During the period of coma the CSF levels of total non-haem iron were significantly raised but levels of chelatable iron were depressed. It may be the loss of intraneuronal iron rather than the transmembrane flux which is important, as iron deficiency in rats leads to a greater susceptibility to this coma (Blake et al, 1985). As we have seen, general anaesthesia and coma are usually associated with disturbances of the glutamate/GABA systems rather than with disturbances of serotonin and noradrenergic systems, which are more related to mood disturbances, cognitive effects and REM sleep. There is, however, some evidence that general anaesthetics affect serotoninergic systems. Halothane reduces 5-HT-induced currents at the 5HT2A receptor (Minami et al, 1997) and general anaesthetics potentiate 5-HT3 receptors (Franks & Lieb, 1997).

11. Thus the coma induced by the combination of desferrioxamine and prochlorperazine may be due to inhibition of the glutamate synapse due to low levels of iron in the post-synaptic neuron, although additional actions at serotoninergic and perhaps other receptors cannot be ruled out. In addition, the side chain of prochlorperazine somewhat resembles a polyamine such as spermidine, and there is a polyamine modulatory site on the NMDAr. This indicates that a possible antagonist effect of prochlorperazine at the polyamine site on the NMDAr may be involved.

12. Since desferrioxamine is hydrophilic, it may be asked how it could cross the cell membrane in the manner required by this hypothesis. Ollinger and Brunk (1995) report that desferrioxamine is taken up by endocytosis in the case of hepatocytes, and is then transported to the acidic endosome. Furthermore, desferrioxamine inhibits the peroxidation and lysis of lysosomal membranes by chelating intralysomal iron. The next sections (paragraphs 9-16) address current knowledge of the biochemical mechanisms that may support the operations of the brain related to consciousness.

13. The general statement that glutamatergic mechanisms may be relevant requires investigation. The details of this system include:

    (i) the redox balance at the glutamate synapse and inside neurons,

    (ii) the role of interactions between the dopamine system and the
    glutamate synapse,

    (iii) the possible role of iron-dopamine complexes,

    (iv) the possible role of endocytosis of receptors and their
    subsequent processing inside the postsynaptic neuron.

14. The first question to be addressed is how intraneuronal levels of iron could affect glutamate synaptic function and so lead to a loss of consciousness. I would suggest the following hypothesis: There is considerable evidence to suggest than one important factor determining the plasticity of the glutamate synapse (growth and diminution of existing synapses, and the supply of new synapses and removal of old ones) is the redox balance at the synapse, both in the synaptic cleft and inside the dendritic spine, which I have reviewed elsewhere (Smythies, 1997a). There is a constantly shifting balance at the glutamate synapse between neurotoxic reactive oxygen species (ROS) and reactive nitrogen species (RNS) on the one hand, and protective antioxidants on the other. The ROS include the superoxide anion, the hydroxyl radical and hydrogen peroxide. These are produced by enzymes in the post-synaptic cascade, such as prostaglandin H synthase and nitric oxide synthase, as well as by mitochondria. RNS include the nitric oxide radical and peroxynitrite. ROS and RNS would tend to lead to spine deletion if produced in excess. However, ROS do not just function as neurotoxins. They may do this when produced in excess, or when antioxidant defenses are inadequate, but normally ROS are important signaling molecules in their own right. There is evidence to suggest that they play such a role in many cellular process such as cell growth, chemotaxis, apoptosis, transcription factor activation, gene expression and others (Suzuki et al, 1997; Kamata and Hirata, 1998). An interesting link between iron, ROS and antioxidants has been provided by Fuchs (1997), who showed that the level of expression of the gene for the important antioxidant enzyme glutathione peroxidase is modulated by intracellular chelatable iron levels, as well as by the level of oxidative stress.

15. Protective antioxidants available at the glutamate synapse include ascorbic acid (vitamin C), the dipeptide carnosine, and glutathione. The effect of positive reinforcement on this system may be mediated by the release of dopamine from its boutons-en-passage, which are attached to the sides of many glutamate synapses (Kotter, 1994). Dopamine exerts powerful anti-oxidant effects through three separate mechanisms: (a) the direct scavenging of ROS and redox cycling between dopamine and dopamine quinone, in conjunction with an antioxidant such as ascorbate or glutathione (Liu and Mori 1994); (b) the activation of D 2 dopamine receptors, which induces the synthesis of an antioxidant enzyme, probably catalase (Sawada et al, 1998) in the post-synaptic neuron; and (c) forming dopamine-iron complexes which act as powerful scavengers of superoxide anions by redox cycling between ferrous and ferric iron and between dopamine and dopamine quinone (Zhao et al, 1998). This system effectively transmutes 2 molecules of superoxide into oxygen and 3 molecules of superoxide into the much less toxic hydrogen peroxide. How do iron and dopamine molecules come into contact, in order that this third mechanism may occur? Since free iron is extremely reactive, almost all the iron in the body is located inside shielding proteins, mainly the iron storage protein ferritin and the iron transport proteins transferrin and lactoferrin. One answer is suggested by the following facts.

16. In cell biology, a paradigm change has recently occurred relating to the mechanism by which receptors for neurotransmitters and neuromodulators function. It used to be thought that when a receptor located in the membrane bound a molecule of the transmitter/modulator it underwent a conformational change. In some cases this opened an ionic channel, while in others it activated a second molecule (such as a G-protein) which started a post-synaptic cascade. The receptor itself was supposed to eject that molecule of transmitter/modulator and then wait in the membrane for the next, when the whole process would be repeated. Although it was recognized that the receptor molecule was eventually replaced, possibly because of accumulated oxidative damage, the general picture of the plasma membrane was that seen under the microscope, i.e., a motionless structure.

17. It is now known that much of the entire membrane, proteins and lipids alike (Kobayashi et al, 1998), is in a constant state of flux, being continually endocytosed, processed by the endosome system and then recycled back to the plasma membrane. This process involves G-protein linked and other similar receptors (Koenig & Edwardson, 1997; Mukherjee et al, 1997). When one of these binds with a molecule from the transmitter/modulator (e.g., a neuropeptide, catecholamine or acetylcholine), the receptor-ligand complex is rapidly endocytosed inside a clathrin-lined pit which converts to a vesicle inside the post-synaptic neuron. This is transported (~10 minutes) to the tubulovesicular endosome system. Here, the vesicle membrane fuses with the membrane of the early endosome and delivers the receptor-ligand complex into the lumen of the endosome, where the acidic environment leads to the dissociation of the ligand from the receptor protein. The ligand is then transported to the late endosome, where the receptor protein is subjected by the endosome/lysosome system to a triage process. This includes those receptor proteins that have been down-regulated by phosphorylation.

18. Because the phosphate groups cannot be removed in situ, the protein is endocytosed to allow their removal, and the newly sensitized receptor is returned to the external membrane (Ferguson & Caron, 1998). Other proteins subject to the triage process may include those damaged by oxidative attack. These have to be broken down into their constituent amino acids. The oxidized amino acids are metabolized and excreted, while the undamaged ones are recycled. The same process may apply to membrane lipids which are also under constant oxidative attack. Dephosphorylated and reduced receptors are recycled back to the cell membrane. Normal proteins and lipids are recycled by the appropriate mechanisms, and endosome membrane is also recycled back to the surface. The whole cycle in some cases takes around 30 minutes (e.g., for NGF receptors (Zapf-Colby and Olefsky,1998)). It seems unlikely that the function of the endocytotic mechanism is only to resensitize receptors, since much of the membrane itself is endocytosed.

19. If the ligand is a polypeptide it is transmitted to the cell nucleus, where it plays an essential role in gene expression (Jans and Hassan, 1998). As Koenig and Edwardson (1997) note in the case of polypeptide transmitters and receptors "the purpose of endocytosis is to capture the ligand for consequent use by the cell". For our present purposes it is of particular interest that dopamine G-protein related receptors are also rapidly and robustly endocytosed following transmitter binding (Dumartin et al, 1998). The D1 receptor is endocytosed by clathrin-lined vesicles, which also transport the iron transporter transferrin, whereas the non-clathrin lined vesicles that endocytose the D2 receptor do not perform this additional function (Vickery et al, 1998). Koenig and Edwardson (1997) state that low affinity agonists (like muscarine or dopamine) are unlikely to be internalized in sufficient quantity to "cause significant receptor activation in endosomes". However, the relationship between internalization and the intrinsic activity of the ligand is non-linear, and so very weak partial agonists can produce significant receptor internalization (Szekeres et al, 1998). Moreover, the role of intracellular dopamine may not be receptor activation but something quite different. This prompts the question of what, if anything, could be the function of dopamine inside the post-synaptic neuron?

20. In the post-dopamine D1 receptor endosome system, transferrin and the D1 receptor-dopamine complex co-localize in the same endosome, which enables chelatable iron and dopamine to come into close contact. The path of iron from the late endosome to its target - the iron-containing enzymes being synthesized in the neuron (including enzymes like tyrosine and tryptophan hydroxylase and certain mitochondrial enzymes) - is not clear. There have been suggestions that a small molecule acts as the carrier (Jacobs 1977; Bradbury 1997). However, when Vyoral and Petrk (1998) used gel electrophoresis, they could not find any evidence for these. They suggested that the endosome is physically in contact with all the structures that use iron (mitochondria, ribosomes, etc.), and that the iron is transported from one to the others by means of an extensive system of channels and tubes.

21. Breuer et al (1997) found that the catalytic potential of iron was highest while in transit between the endosomes and cytosolic ligands. They also found that iron is released from endosomes and enters a cytoplasmic pool at a concentration of 0.3-0.5 mM (Breuer et al,1995). The mean transit time through the chelatable pool is 1-2 hours. Moos and Morgan (1998) have recently presented experimental evidence for the existence of low-molecular weight transporters for iron in the brain and in cerebrospinal fluid, and suggest that citrate or ascorbate might act in this way. If the postulated iron-dopamine complex is carried attached to some protein, this may perhaps explain Vyoral and Petrk's results. The function of dopamine inside the post-synaptic neuron, in the form of an iron-dopamine complex, may be similar to what I have suggested is its role in the synapse; that is, as an antioxidant protection for the spine and dendrite. This would have the added advantage of the safe transport of iron to its cytoplasmic destinations.

22. As noted earlier, free iron is far too toxic to be loose in the cytoplasm in anything more than minute quantities. The dopamine complex-mediated iron transporter could be transported to the sites of iron usage through the cytoplasm, or within extensions of the endoplasmic tubules, as proposed by Vyoral and Petrk. In the first case the superoxide anions would encounter the dopamine-iron complexes in the cytoplasm, and in the second case superoxide anions could enter the endosomes from their main sites of production i.e., the mitochondria, to which the endoplasmic tubes would give direct access. Qian et al (1997) have reviewed the whole question of how iron is transported from inside the endosome to its cytosolic targets. Their conclusion is that little is known about this subject, and that there may be multiple carriers (e.g., p97, integrin, H+-ATPase and the transferrin receptor).

23. There may also be different transporters in different cells, and the evidence suggests some type of carrier 'chaperone' protein rather than an iron-conducting channel. If the dopamine-iron complex forms in the endosome, it may be transmitted to the cytosol by such a carrier mechanism. Alternatively, there may be separate transport mechanisms for iron and dopamine out of the endosome and the complex forms only in the cytosol. Qian et al (1998) state that iron is maintained in a chelateable pool in the cytoplasm after leaving the endosome. Catecholamine-iron complexes function as iron siderophores for the bacterium Listeria monocytogenes, acquiring iron from the environment and transporting it into the cell by means of a ferric reductase in the membrane (Coulanges et al, 1997). Perhaps the endosome membrane has a similar ferric reductase system. Mitochondria, descended from bacteria, may act in the same way. It is also possible that the dismuting dopamine-iron complex forms in the glutamatergic synaptic cleft itself. It is likely that free dopamine, and free iron in low concentrations, are present in this location. Superoxide anions could only be present if the neuronal membrane contains superoxide channels, as the erythrocyte does. No information on this point is currently available.

24. This system enables us to explain how the combination of desferrioxamine and prochlorperazine induce coma. One factor may be that in promoting iron transport across membranes, this combination depletes intracellular iron by the synergistic effect of its components. This prevents the formation of the dopamine-iron complex inside the post-synaptic cell at the glutamate synapse, so that the superoxide ions, produced by mitochondria and enzymes on the post NMDAr cascade, cannot be scavenged. This effectively shuts down the glutamate synapse and the functional effect is akin to the effect of ketamine, although the details are different, as ketamine acts at the receptor level, while the combination of desferrioxamine and prochlorperazine may act at the post-synaptic level. A second factor, or contributory mechanism, may be that prochlorperazine (the side chain of which has structural similarities with spermine-type polyamines) blocks the polyamine site on the NMDA receptor protein, and thus acts in a synergistic manner, with lowered intracellular iron leading to down regulation of the NMDA receptor.

25. Thus, it could be argued that consciousness depends on the general level of activity of the whole cortex via a wide range of cortico-thalamic relays - as determined by a proper balance between glutamatergic activation and GABAergic inhibition, plus modulation by other factors, such as local cholinergic neurotransmission - as much as on the activity of strategically placed brain stem nuclei such as the pedunculopontine nucleus and the intralaminar nuclei of the thalamus (Smythies, 1997b). However, it is not clear why it is NMDAr antagonists rather than AMPAr or mGlur antagonists that induce coma. This would suggest a further field for study.

26. This hypothesis mirrors that reached by Angel (1991), that anaesthesia is associated with a general impairment of communication between the thalamus and cortex. Consciousness may be realized only when some critical threshold of global neural activity is present. Llins and Par (1991) have suggested that the content of consciousness is provided by activity in cortico-thalamic loops which involves specific thalamic nuclei, and that the 'binding' of these into a 'unitary experience' involves activity in the intralaminar thalamic nuclei. However, another hypothesis is possible. As the molecular targets of general anaesthetics (in either the classical or the receptor-based theories) are widely distributed in the brain, it seems likely that anaesthesia would affect all brain systems in a rather unselective manner. What may be of greater significance is their interaction with a subclass of key 'consciousness neurons' (should these exist), rather than their global effect on the brain. Thus it can be argued that general anaesthetics can tell us little or nothing about the specific neural constituents of conscious phenomena. It may be possible to clarify this point through the following experimental approach:

    (i) prepare a number of rats with indwelling canulae suitable for
    local perfusion in a number of key locations, such as the medial
    thalamus, midbrain tegmental area, midbrain reticular formation,
    and basal forebrain,

    (ii) locally perfuse these awake animals with a series of general
    anaesthetic agents in these different anatomical loci and see if
    coma develops.

27. Clearly, if a full coma (general anaesthesia) followed such a perfusion of area A but not of areas B, C, and D, this would indicate that the general anaesthetic effect of that agent was mediated by its effects on area A, as a necessary and sufficient cause of the coma, and not by any effects on areas B, C, or D, or on the cortex as a whole. Similarly, if no such local effect could be found to support the hypothesis, then general anaesthesia must depend on more widespread effects. However, it could not be assumed from such a result alone that a widespread cortical inhibition is required. Additional experiments involving the local perfusion of combinations of these subcortical areas would first be necessary. It would not seem technically possible to perfuse the entire cortex without leakage to subcortical areas, or to perfuse the reticular nucleus of the thalamus uniquely, owing to its special anatomy. Gases such as halothane and nitrous oxide might produce technical difficulties but agents such as anaesthetic barbiturates, ketamine and the desferrioxamine/prochlorperazine mixture should not. I can find no evidence, in spite of an intensive literature search, that this experiment has ever been done.

28. In the past, the generally accepted position was that the key neuronal functions to be considered for any aspect of brain function (behavioral as well as pharmacological) were neurotransmitter synthesis, release, activation of specific postsynaptic (and some presynaptic) receptors, and a complex series of postsynaptic cascades involving various nucleotides. These provided the framework for investigating what were considered the important events at this level of brain function, events that might relate, inter alia, to the evolution of consciousness. Recent advances in cell biology (which I mentioned in section 12, but with which many neuroscientists seem unfamiliar) have added a new and highly significant dimension to this research.

29. This new territory demonstrates that neurotransmitter receptors are subject to a continuing and massive process of endocytosis into the post-synaptic neuron, transfer to the endosome system (in some cases in clathrin coated vesicles and in other cases in non-clathrin coated vesicles), molecular processing in the endosome system, and a final triage process, during which some of the receptor protein is broken down by lysosomes but most is recycled back to the external membrane. In many cases the bound neurotransmitter (or neuromodulator) is endocytosed together with the receptor. This means that agents like general anaesthetics, which appear to affect a number of receptors, may be affecting the endocytotic mechanism common to these receptors. It also entails the revolutionary concept that some neurotransmitters, or neuromodulators, could play a key role inside the post-synaptic neuron. In the case of neuropeptide neuromodulators, it is now known that they are trafficked to the nucleus, where they play a role in transcription processes. No such role has as yet been discovered for smaller neurotransmitters. I presented an hypothesis based on this possibility relating to dopamine earlier (paragraph 15). At present there is no direct evidence linking the endocytotic mechanism to the maintenance of consciousness or the induction of coma. However, this revolution in neuronal biology is so important that it deserves inclusion in any consideration of any higher neuronal activity.


30. Another instance of coma is familiar non-REM (NREM) sleep. During both the awake state and REM sleep we are basically conscious, even though the phenomenology of these states is very different. Recently two good reviews of this topic have appeared (Kahn et al, 1997; Hobson et al, 1998). The key feature of the awake state appears to be a balanced activity in the cholinergic, noradrenergic and serotoninergic systems permitting attentive behavior. During REM sleep, cholinergic activity in the peribrachial pedunculopontine and lateral dorsal tegmental cholinergic nuclei increases, and activity in the noradrenergic and serotoninergic activity greatly decreases. In NREM sleep, activity of all three of these neurotransmitters is low and there is deactivation of almost all brain systems. This concurs with the theory that general anaesthesia is associated with a widespread down regulation of the cortex. In NREM sleep there is also a reduction in the activity of some peribrachial cholinergic cells, which results in a change from tonic to phasic firing of thalamic relay nuclei. This is consistent with the specific effect of general anaesthetics in reducing cholinergic activity as reviewed above (see section 4 (i)). Kahn et al (1997) conclude that "conscious actions require activated cortical networks".

NOTE: I have written about the problem of the nature of consciousness elsewhere (Smythies 1994 a, 6).


Albrecht, J. (1998) 'Roles of neuroactive amino acids in ammonia toxicity.' Neuroscience Research 51: 133-138.

Andoh, T., Furuya, R., Oka, K., Hattori, S., Watanabe, I., Kamiya, Y. & Okumura, F. (1997) 'Differential effects of thiopental on neuronal nicotinic acetylcholine receptors and P2x purinergic receptors in PC 12 cells.' Anesthesiology 87: 1199-1209.

Andrade, J. (1996) 'Investigations of hypethesia: using anesthetics to explore the relationships between consciousness, learning and memory.' Consciousness and Cognition 5: 562-580

Angel, A. (1991) 'Adventures in anesthesia.' Experimental Physiology 76: 1-38.

Blake, D. R., Winyard. P., Lunec. J., Williams, A., Good, P. A., Crewes, S. J., Gutteridge, J. M. C., Rowley, D., Halliwell, B., Cornish, A. & Hider, R. C. (1985) 'Cellular and ocular toxicity induced by desferrioxamine.' Quarterly Journal of Medicine 56: 345-355.

Breuer, W., Epsztejn, S. & Cabantchik, Z.L. (1995) 'Iron acquired from transferrin by K562 cells is delivered into a cytoplasmic pool of chelatable iron (II).' Journal of Biological Chemistry 270: 24,209-24,215.

Breuer, W., Greenberg, E. & Cabantchik, Z. I. (1997) 'Newly delivered transferrin iron and oxidative cell injury.' FEBS Letters 403: 213-219.

Bradbury, M. W. B. (1997) 'Transport of iron in the blood-brain-cerebrospinal fluid system.' Journal of Neurochemistry 69: 443-454.

Bullock, R., Zauner, A., Woodward, J.J., Myseros, J., Choi, S.C., Ward, J.D., Marmarou, A. & Young, H.F. (1998) 'Factors affecting excitatory amino acid release following severe human head injury.' Journal of Neurosurgery 89: 507-518.

Coulanges, V., Andre, P., Ziegler, O., Buchheit, L. & Vidon, D. J-M. (1997). 'Utilization of iron-catecholamine complexes involving ferric reductase activity.' Infection and Immunity 65, 2778-2785.

Crick, F. (1994) 'The Astonishing Hypothesis. New York, Scribners. Denson, D. D., Worrell, R. T. & Eaton, D. C. (1996) A possible role for phospholipase A2 in the action of general anesthetics.' American Journal of Physiology 270: C636-C644.

Dumartin, B., Caill, I., Gonon, F. & Bloch, B. (1998) 'Internalization of D1 dopamine receptor in striatal neurons in vivo as evidence of activation by dopamine agonists.' Journal of Neuroscience 18: 1650-1661.

Dzoljic, M. & van Duijn, B. (1998) 'Nitrous oxide-induced enhancement of g-aminobutyric acid-mediated chloride currents in acutely dissociated hippocampus neurons.' Anesthesiology 88: 473-480.

Eckenhoff, R. G. & Johansson, J. S. (1997) 'Molecular interactions between inhaled anesthetics and proteins.' Pharmacological Reviews 49: 343-367.

Ferguson, S. S. & Caron, M. G. (1998) 'G protein-coupled receptor adaptation mechanisms.' Seminars in Cellular and Developmental Biology 9: 119-127.

Fink, K., Zentner, J. & Gthert, M. (1994) 'Increased GABA release in the human brain cortex as a potential pathogenetic basis of hyperosmolar coma.' Journal of Neurochemistry 62: 1476-1481.

Fiset, P., Paus, T., Daloze, T., Plourde, G., Meuret, P., Bonhomme, V., Hajj-Ali, N., Backman, S.B., & Evans, A.C. (1999) 'Brain mechanisms of propofol-induced loss of consciousness in humans: a positron emission tomographic study.' Journal of Neuroscience 19: 5506-5513.

Flohr, H. (1995) 'Sensations and brain processes.' Behavioural Brain Research. 71: 157-161.

Franks, N. P. & Lieb, W. R. (1994) 'Molecular and cellular mechanisms of general anaesthesia.' Nature 367: 607-614.

Franks, N., P. & Lieb, W. R. (1997) 'Anesthetics set their sites on ion channels.' Nature 389:334-335.

Franks, N. P. & Lieb, W. R. (1998) 'A serious target for laughing gas.' Nature. Medicine 4: 383-384.

Fuchs, O. (1997) 'Effects of intracellular chelatable iron and oxidative stress on transcription of classical cellular glutathione peroxidase gene in murine erythroleukemia cells.' Neoplasma 44: 184-191.

Griffiths, R., Grieff, J.M.C., Boyle, E., Rowbottom, D.J. & Norman, R. I. (1994) 'Volatile anesthetic agents inhibit choline uptake into rat synaptosomes.' Anesthesiology 81: 953-958.

Hashimoto, Y., Hashimoto, Y., Hirota, K. & Matsuki, A. (1998) 'Inhibited hypothalamic histamine metabolism during isoflurane and sevoflurane anaesthesia in rats.' Acta Anaesthesologica Scandinavica 42: 858-863.

Hobson, J. A., Stickgold, R. & Pace-Schott, E. F. (1998) 'The neuropsychology of REM sleep dreaming.' NeuroReport 9: R1-R14.

Jacobs, A. (1977) 'Low molecular weight intracellular iron transport compounds.' Blood 50: 433-439.

Janoff, A. S, Pringle, M. J. & Miller, K. W. (1981) 'Correlation of general anesthetic properties with solubility in membranes.' Biochimica et Biophysica Acta 649: 125-128.

Jans, D. A. & Hassan, G. (1998) 'Nuclear targeting by growth factors, cytokines, and their receptors: a role in signaling?' Bioessays 20: 400-411.

Kabatnik, M., Heist, M., Beiderlinden, K. & Peters, J. (1999) 'Hepatic encephalopathy-a physostigmine-reactive central anticholinergic syndrome?' European Journal of Anaesthesiology 16: 140-142.

Kahn, D., Pace-Schott, E. F. & Hobson, J.A. (1997) 'Consciousness in waking and dreaming: the roles of neuronal oscillation and neuromodulation in determining similarities and differences.' Neuroscience 78: 13-38.

Kamata, H. & Hirata, H. (1998) 'Redox regulation of cellular signaling.' Cellular Signaling 11: 1-14.

Katayama, Y., Mori, T., Maeda, T. & Kawamata, T. (1998) 'Pathogenesis of the mass effect of cerebral contusions: rapid increase in osmolarity within the contusion necrosis.' Acta Neurochirurgica. Supplementum 71: 289-292.

Keifer, J. C., Baghdogan, H. A. & Lydic, R. (1996) 'Pontine cholinergic mechanisms modulate the cortical electroencephalographic spindles of halothane anesthesia.' Anesthesiology 84: 945-954.

Kobayashi, T., Gu, F. & Gruenberg, J. (1998) 'Lipids, lipid domains and lipid-protein interactions in endocytic membrane traffic.' Seminars in Cell Development and Biology, 9: 51-526.

Koenig, J. A. & Edwardson, J. M. (1997) 'Endocytosis and recycling of G protein-coupled receptors.' Trends in Pharmacological Sciences 18: 276-287.

Kotter, R. (1994) 'Postsynaptic integration of glutamatergic and dopaminergic signals in the striatum.' Progress in Neurobiology 44: 163-196.

Kulkarni, R. S., Zorn, L. J., Anantharam, V., Bayley, H. & Triestman, S. N. (1996) 'Inhibitory effects of ketamine and halothane on recombinant potassium channels from mammalian brain.' Anesthesiology 84: 900-909.

Kurihara, J., Hu, B. R. & Wieloch, T. (1994) 'Changes in the tyrosine phosphorylation of mitogen-activated protein kinase in the rat hippocampus during and following severe hypoglycemia.' Journal of Neurochemistry 63: 2346-2348.

Liachenko, S., Tang, P., Somogyi, G. T. & Xu, Y. (1998) 'Comparison of anesthetic and non-anesthetic effects on depolarization-evoked glutamate and GABA release from mouse cerebrocortical slices.' British Journal of Pharmacology 123: 1274-1280.

Lin, L., Koblin, D. D. & Wang, H. H. (1995) 'Effects of halothane on the nicotinic acetylcholine receptor from Torpedo Californica.' Biochemical Pharmacology 89: 1085-1089.

Liu, J. & Mori, A. (1994) 'Monoamine metabolism provides an antioxidant defense in the brain against oxidant- and free radical-induced damage.' Archives of Biochemistry and Biophysics 302: 118-127.

Liu, Y., Dilger, J. P. & Vidal, A. M. (1994) 'Affects of alcohols and volatile anesthetics on the activation of nicotinic acetylcholine receptor channels.' Molecular Pharmacology 45: 1235-1241.

Llins, R.R. & Par, D. (1991) 'Of dreaming and wakefulness.' Neuroscience 44; 521-535.

Lopez, M. M. & Kosk-Kosicka, D. (1996) 'How do volatile anesthetics inhibit CA(2+)-ATPases?' Journal of Biological Chemistry 270: 28,239-28,245.

Mills, C. D., Robertson, C. S., Contant, C. F. & Henley, C. M. (1997) 'Effects of anesthesia on polyamine metabolism and water content in the rat brain.' Journal of Neurotrauma 14: 943-949.

Minami, K., Vanderak, T. W., Minami, M. & Harris, R.A. (1997) 'Inhibitory effects of anesthetics and ethanol on muscarinic receptors expressed in Xenopus oocytes.' European Journal of Pharmacology 339: 237-243.

Moos, T. & Morgan, E. H. (1998) 'Evidence for low molecular weight non-transferrin-bound iron in rat brain and cerebrospinal fluid.' Journal of Neuroscience Research 54: 486-494.

Mori, T., Kawamata, T., Katayama, Y., Maeda, T., Aoyama, N., Kikuchi, T. & Uwahodo, Y. (1998) 'Antioxidant, OPC-14117, attenuates edema formation, and subsequent tissue damage following cortical contusion in rats.' Acta Neurochirurgica. Supplementum. 71: 120-122.

Mukherjee, S., Ghosh, R. N. & Maxfield, F. R. (1997) 'Endocytosis.' Physiological Reviews 77: 759-803.

Ollinger, K. & Brunk, U. T. (1995) 'Cellular injury induced by oxidative stress is mediated through lysosomal damage.' Free Radical Biology and Medicine 19: 565-574.

Qian, Z. M, Tang, P. L. & Wang, Q. (1997) 'Iron crosses the endosomal membrane by a carrier-mediated process.' Progress in Biophyics and Molecular Biology 67: 1-15.

Raines, D. E., Rankin, S. E., & Miller, K. W. (1995) 'General anesthetics modify the kinetics of nicotinic acetylcholine receptor desensitization at clinically relevant concentrations.' Anesthesiology 82: 276-287.

Ratnakumari, L. & Hemmings, H. C. Jr. (1998) 'Inhibition of presynaptic sodium channels by halothane.' Anesthesiology 88: 1043-1054.

Reis, C.R. & Puil, E. (1999a) 'Mechanism of anesthesia revealed by shunting actions of isoflurane on thalamocortical neurons.' Journal of Neurophysiology 81: 1795-1801.

Reis, C.R., & Puil, E. (1999b) 'Ionic mechanism of isoflurane actions on thalamocortical neurons.' Journal of Neurophysiology 81: 1802-1809.

Sawada, H., Ibim, M., Kiharam, T., Urushitanim, M., Akaikem, A., Kimurama, J. & Shimohama, S. (1998) 'Dopamine D2-type agonists protect mesencephalic neurons from glutamate neurotoxicity: mechanisms of neuroprotective treatment against oxidative stress.' Annals of Neurology 44: 110-119.

Scheller, M., Bufler, J., Schneck, H., Kochs, E. & Franke, C. (1997) 'Isoflurane and seroflurane interact with the nicotinic acetylcholine receptor channels in micromolar concentrations.' Anesthesiology 86: 118-127.

Shichino, T., Murakawa, M., Adachi, T., Arai, T., Miyazaki, Y. & Mori, K. (1998) 'Effects of inhalation anesthetics on the release of acetylcholine in the rat cerebral cortex in vivo.' British Journal of Anesthesiology 80: 365-371.

Smythies, J. (1994a) The Walls of Plato's Cave. Aldershot, Avebury Press.

Smythies, J. (1994b) 'Requiem for the Identity theory.' Inquiry 37: 311-329.

Smythies, J. (1997a) 'The biochemical basis of synaptic plasticity and neurocomputation: a new theory.' Proceedings of the Royal Society of London. B. 264: 575-579.

Smythies, J. (1997b) 'The functional neuroanatomy of awareness: with a focus on the role of various anatomical systems in the control of intermodal attention.' Consciousness and Cognition 6: 455-481.

Suzuki, Y. J., Forman, H. J. & Seranian, A. (1997) 'Oxidants as stimulators of signal transduction.' Free Radical Biology and Medicine 22: 269-285.

Szekeres, P. G., Koenig, J. A. & Edwardson, J. M. (1998) 'The relationship between agonist intrinsic activity and the rate of endocytosis of muscarinic receptors in a human neuroblastoma line.' Molecular Pharmacology 53: 759-763.

Tennnigkeit, F., Ries, C. R., Schwarz, D. W. & Puil, E. (1997) 'Isoflurane attenuates resonant responses of auditory thalamic neurons.' Journal of Neurophysiology 78: 591-596.

Vickery, R. G, Christine, C. & von Zastrow, M. (1998) 'Subtype-specific differences in dopamine receptor endocytosis.' Abstracts of the Society for Neuroscience 24: 23.

Vyoral, D. & Petrk, J. (1998) 'Iron transport in K562 cells: a kinetic study using native gel electrophoresis and 59 Fe autoradiography.' Biochimica et Biophysica Acta 1403: 179-188.

Wakamori, M., Ikemoto, Y. & Yamashita, M. (1998) 'Halothane increases the open probability of glycine-activated channel current in rat central neurones.' British Journal of Anaesthesia 80: 840-842.

Yunoki, M., Kawauchi, M., Ukita, N., Noguchi, T., Nishio, S., Ono, Y., Asari, S., Ohmoto, T., Asanuma, M. & Ogawa, N. (1998) 'Effects of lecithinized SOD on sequential changes in SOD activity after cerebral contusion in rats.' Acta Neurochirurgica. Supplementum 71: 142-145.

Zapf-Colby, A. & Olefsky, J. M. (1998) 'Nerve growth factor processing and trafficking events following Trk-A-mediated endocytosis.' Endocrinology 139: 3232-3240.

Zhao, Z. S., Khan, S. & O'Brien, P. J. (1998) 'Catecholic iron complexes as cytoprotective superoxide scavengers against hypoxia:reoxygenation injury in isolated hepatocytes.' Biochemical Pharmacology 56: 825-830.

Zou, Z., De Vente, J. & Johns, R. A. (1996) 'Halothane and isoflurane dose-dependently inhibit the cyclic GMP increase caused by N-methyl-D-aspartate in rat cerebellum: novel localization and quantitation by in vitro autoradiography.' Neuroscience 74: 1069-1075.

Volume: 10 (next, prev) Issue: 026 (next, prev) Article: 1 (next prev first) Alternate versions: ASCII Summary